Woodall George M
National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, NCEA-RTP (B243-01), Research Triangle Park, NC, USA.
Toxicol Appl Pharmacol. 2008 Nov 15;233(1):14-6. doi: 10.1016/j.taap.2007.12.039. Epub 2008 Jul 11.
Risk assessment for human health effects often depends on evaluation of toxicological literature from a variety of sources. Risk assessors have limited resources for obtaining raw data, performing follow-on analyses or initiating new studies. These constraints must be balanced against a need to improve scientific credibility through improved statistical and analytical methods that optimize the use of available information. Computerized databases are used in toxicological risk assessment both for storing data and performing predictive analyses. Many systems provide primarily either bibliographic information or summary factual data from toxicological studies; few provide adequate information to allow application of dose-response models. The Exposure-Response database (ERDB) described here fills this gap by allowing entry of sufficiently detailed information on experimental design and results for each study, while limiting data entry to the most relevant. ERDB was designed to contain information from the open literature to support dose-response assessment and allow a high level of automation in performance of various types of dose-response analyses. Specifically, ERDB supports emerging analytical approaches for dose-response assessment, while accommodating the diverse nature of published literature. Exposure and response data are accessible in a relational multi-table design, with closely controlled standard fields for recording values and free-text fields to describe unique aspects of the study. Additional comparative analyses are made possible through summary tables and graphic representations of the data contained within ERDB.
对人类健康影响的风险评估通常取决于对来自各种来源的毒理学文献的评估。风险评估人员获取原始数据、进行后续分析或开展新研究的资源有限。必须在这些限制与通过改进统计和分析方法提高科学可信度的需求之间取得平衡，这些方法可优化可用信息的利用。计算机化数据库在毒理学风险评估中既用于存储数据，也用于进行预测分析。许多系统主要提供毒理学研究的书目信息或简要事实数据；很少有系统能提供足够信息以应用剂量反应模型。这里描述的暴露-反应数据库（ERDB）通过允许输入每项研究的实验设计和结果的足够详细信息，同时将数据输入限制在最相关的信息上，填补了这一空白。ERDB旨在包含来自公开文献的信息，以支持剂量反应评估，并在进行各种类型的剂量反应分析时实现高度自动化。具体而言，ERDB支持剂量反应评估的新兴分析方法，同时适应已发表文献的多样性。暴露和反应数据可通过关系多表设计获取，有严格控制的标准字段用于记录值，还有自由文本字段用于描述研究的独特方面。通过ERDB中包含的数据的汇总表和图形表示，可以进行额外的比较分析。
